There are ample data indicating that substance abuse is a major societal problem, that female substance abusers have been understudied, and that there appears to be a link between substance abuse and victimization experiences such as rape, sexual molestation, aggravated assault, and other forms of physical assault. Other evidence suggests that the prevalence of victimization and other traumatic events is high, particularly among young women. Still other data indicate that demographic characteristics, family history of substance abuse, the personality trait of sensation seeking, and current anxiety or psychological distress are risk factors for substance use and/or abuse. Given the possible complex relationships among these variables, it is clear that adequate study of casual links between risk factors and substance abuse problems requires the use of a multivariate, longitudinal research design using a large non-helpseeking sample. The major objective of the proposed project is to gather longitudinal data from a reasonably representative sample of adult women that will enable us to test a multivariate risk factor model for substance use and abuse problems among women. A representative national household sample of 4,000 adult women will be located by random-digit-dialing telephone survey methodology and interviewed in three assessment waves. During Wave 1 of the survey, data will be gathered about the following potential risk factors for substance abuse problems: a) demographic characteristics, b) family history of substance abuse, c) the personality trait of sensation seeking, d) victimization history, and e) history of other traumatic life events. Data will also be gathered about current psychological stress and distress, respondent's attributions of vulnerability, control and helplessness, and about current substance use and abuse. During Wave 2, to be conducted one year after Wave 1, data will be collected about current psychological stress and distress, current substance use and abuse, and about new incidents of victimization and other traumatic events that occurred since Wave 1. The Wave 3 interview, to be conducted two years after Wave 1, will be identical to that used in Wave 2. A variety of multivariate data analytic techniques will be used to test hypothesized relationships among major variables of interest. Structural modeling procedures will be used to test causal hypotheses using the longitudinal data.